Re: Logging NDB and SQL... endless loop issue?

On 18-Feb-2018 10:50 -0700, Bradley Stone wrote:

On Sun, Feb 18, 2018 at 10:06 AM, Bradley Stone wrote:

I am looking into putting together some simple NDB and SQL logging
for a customer using the appropriate exit points.

The issue I can see is, if both exit points are working, and the
logging goes to normal table, this seems like it would cause an
infinite loop of logging, no matter if RLA or SQL is used to write
the information. Unless I'm overthinking it.:)

Ie... an SQL statement comes in Exit program gets it, and writes a
log of it to a table (using SQL) ..... this causes the exit program
be called again...for each record itself writes

If we use RLA to write the SQL log, and have NDB logging, same
thing. SQL statement comes in SQL exit point logs using RLA, NDB
exit is called NDB exit writes to log and any way it does causes
either NDB or SQL pme PF the exit points will be called again.
<<SNIP>>

Also, I do understand that I could write the logs using write, and
then hardcode in the NDB exit point that if it's this particular
file to NOT log a record, but I don't want to hardcode and I can
easily forsee if it is a setting somewhere instead that it would get
changed "accidentally" causing issues.

Same for if SQL was used to write the log data... in the SQL exit
point code a "don't log" for the specific file(s). Same issue could
occur.

Maybe it will make more sense once I see the actual data coming
through.. just thinking out loud in a cloud of fever dementia right
now. .haha..

Seems to me, that the intended "logging" will most likely occur in the job of the exit point processing [of the SQL or NDB activity], and thus the "write" activity by the exit program would be *neither* of an SQL nor an NDB request that would be subject to the logging.? That's typical of how those exits would be utilized [and thus unaffected by the perceived-as potential issue].?